Open Imaging Finding Models

A Python workspace for managing Open Imaging Finding Models - structured data models used to describe medical imaging findings in radiology reports.

Packages

Package	Description	Install
findingmodel	Core models, Index API, MCP server, findingmodel CLI	pip install findingmodel
findingmodel-ai	AI-powered tools, findingmodel-ai CLI	pip install findingmodel-ai
anatomic-locations	Anatomic location queries, anatomic-locations CLI	pip install anatomic-locations

Internal packages (not published):

• oidm-common - Shared infrastructure
• oidm-maintenance - Database build/publish (maintainers only)

Quick Start

# Install core package
pip install findingmodel

# Install AI tools (requires API key)
pip install findingmodel-ai

Create a .env file:

# Required for embedding-based search and AI tools
OPENAI_API_KEY=your_key_here

# Or use another provider
ANTHROPIC_API_KEY=your_key_here
GOOGLE_API_KEY=your_key_here

Using the Index

import asyncio
from findingmodel import Index

async def main():
    async with Index() as index:
        # Search for finding models
        results = await index.search("pneumothorax", limit=5)
        for result in results:
            print(f"- {result.name}: {result.description}")

asyncio.run(main())

Using AI Tools

import asyncio
from findingmodel_ai.authoring import create_info_from_name

async def main():
    info = await create_info_from_name("pneumothorax")
    print(f"Name: {info.name}")
    print(f"Description: {info.description}")

asyncio.run(main())

CLI Tools

# Core tools
findingmodel config              # View configuration
findingmodel stats               # Index statistics
findingmodel search "pneumothorax"  # Search finding models

# AI tools
findingmodel-ai make-info "pneumothorax"     # Generate finding info
findingmodel-ai make-stub-model "finding"    # Create model template

# Anatomic locations
anatomic-locations search "knee joint"       # Search locations

Configuration

See Configuration Guide for:

• AI provider setup (OpenAI, Anthropic, Google, Ollama)
• Model tier configuration
• Per-agent model overrides
• Database path customization

Documentation

• Configuration Guide - API keys, providers, model selection
• Anatomic Locations Guide - Anatomic location features
• MCP Server Guide - Claude Desktop integration
• Database Management - For maintainers

Development

# Install all packages in development mode
uv sync

# Run tests
task test           # Unit tests (no API calls)
task test-full      # Integration tests

# Run checks
task check          # Format, lint, type check

License

MIT